Intelligent sampling and quality inspection system and method for grain depot

Abstract

The invention discloses a grain depot intelligent sampling quality inspection system and method, and relates to the technical field of intelligent grain depot detection, and the method comprises the steps: collecting environment data, three-dimensional point cloud data and surface acoustic wave group velocity data, and carrying out the inversion calculation through a Love wave equation, and obtaining a three-dimensional density distribution diagram; calculating the density gradient in the vertical direction according to the three-dimensional density distribution diagram to obtain a stratified sampling parameter table; according to the stratified sampling parameter table, stratified puncturing is conducted on the grain pile, actual sampling depth data are generated, and a three-way sample valve is used for conducting sample distribution detection on the actual sampling depth data. Through layered detection of the three-way sample valve and near-infrared X-ray combined analysis, accurate detection of moisture, protein and heavy metal content is realized, and a detection result is matched with a density distribution diagram through a constrained Kriging interpolation method to generate a quality safety index.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent grain depot detection, in particular to a grain depot intelligent sampling and quality inspection system and method. Background Art

[0002] Traditional sampling methods lack accurate measurement of the density distribution inside the grain pile, resulting in insufficient representativeness of the samples and inability to comprehensively reflect the quality of the grain. Secondly, existing technologies mostly adopt single detection means, making it difficult to simultaneously obtain the physical properties (such as density) and chemical components (such as moisture, protein, and heavy metal content) of the grain, which limits the comprehensiveness and accuracy of the detection results.

[0003] Existing technologies have not effectively integrated surface acoustic wave group velocity data with three-dimensional point cloud data and environmental data, resulting in limited accuracy and efficiency of density distribution calculation. In addition, existing technologies mostly rely on single detection means and lack the ability to comprehensively evaluate the quality and safety indicators of the grain. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a grain depot intelligent sampling and quality inspection system and method, which solves the problems of insufficient accuracy of density distribution calculation and lack of comprehensive evaluation ability of grain quality and safety indicators in the prior art.

[0006] To solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a grain depot intelligent sampling and quality inspection method, which includes collecting environmental data, three-dimensional point cloud data, and surface acoustic wave group velocity data, and using the Love wave equation for inversion calculation to obtain a three-dimensional density distribution map;

[0008] Calculating the vertical density gradient based on the three-dimensional density distribution map to obtain a stratified sampling parameter table;

[0009] Stratified piercing the grain pile according to the stratified sampling parameter table to generate actual sampling depth data, and using a three-way sampling valve to perform sample distribution detection on the actual sampling depth data;

[0010] Adopting near-infrared spectroscopy and X-ray fluorescence to analyze the moisture, protein, and heavy metal content in the grain pile, and matching the analysis results with the three-dimensional density distribution map to obtain quality and safety indicators.

[0011] As a preferred solution of the grain depot intelligent sampling and quality inspection system and method of the present invention, wherein: collecting environmental data, three-dimensional point cloud data, and surface acoustic wave group velocity data includes the following steps,

[0012] Scanning the surface of the grain pile with a lidar to generate three-dimensional point cloud data;

[0013] On the surface sensor array of the grain pile, by emitting surface acoustic waves and measuring the propagation time difference between adjacent sensors, the group velocity data of the surface acoustic waves is obtained.

[0014] As a preferred embodiment of the intelligent sampling and quality inspection system and method for grain depots of the present invention, wherein: the inversion calculation using the Love wave equation to obtain the three-dimensional density distribution map includes the following steps.

[0015] Based on the surface acoustic wave group velocity data, use the Love wave equation to invert the grain pile density and generate a three-dimensional density distribution.

[0016] Use the grid interpolation combination method to combine the three-dimensional density distribution with the three-dimensional point cloud data to obtain a three-dimensional density distribution map.

[0017] As a preferred embodiment of the intelligent sampling and quality inspection system and method for grain depots of the present invention, wherein: calculating the vertical density gradient according to the three-dimensional density distribution map to obtain the stratified sampling parameter table includes the following steps.

[0018] Align the density data with the lidar point cloud coordinates through the grid alignment method, and use the central difference method to calculate the density gradient layer by layer in the vertical direction of the grain pile.

[0019] Set the density mutation layer threshold based on the density gradient layer by layer in the vertical direction of the grain pile. When the density gradient is greater than the density mutation layer threshold, divide the boundary edge of the grain pile.

[0020] Divide the grain pile into N layers according to the boundary edge of the grain pile to obtain the stratified sampling grain pile parameter table.

[0021] As a preferred embodiment of the intelligent sampling and quality inspection system and method for grain depots of the present invention, wherein: performing stratified piercing on the grain pile according to the stratified sampling parameter table to generate actual sampling depth data includes the following steps.

[0022] Perform priority sorting on the stratified sampling parameter table through the max heap data structure, dynamically analyze the high-risk areas, and dynamically adjust the piercing speed in descending order of the density gradient intensity.

[0023] Use a Terfenol-D magnetostrictive actuator to vertically pierce into the grain pile at the dynamically adjusted piercing speed, and record the actual sampling depth data in real time.

[0024] As a preferred embodiment of the intelligent sampling and quality inspection system and method for grain depots of the present invention, wherein: using a three-way sampling valve to perform sample distribution detection on the actual sampling depth data includes the following steps.

[0025] Set the surface humidity threshold and deep humidity threshold of the grain pile according to the overall humidity of the grain pile collected.

[0026] Use a three-way sample valve to perform sample distribution detection on the actual sampling depth data;

[0027] When the actual sampling depth data is less than the surface humidity threshold, perform mold detection;

[0028] When the surface humidity threshold is less than or equal to the actual sampling depth data and the actual sampling depth data is less than or equal to the deep layer humidity threshold, perform moisture detection;

[0029] When the actual sampling depth data is greater than the deep layer humidity threshold, perform heavy metal detection.

[0030] As a preferred solution of the grain depot intelligent sampling quality inspection system and method described in the present invention, wherein: use near-infrared spectroscopy and X-ray fluorescence to analyze the moisture, protein and heavy metal content in the grain pile, and match the analysis results with the density distribution map to obtain the quality safety indicators, including the following steps,

[0031] Based on the stratification parameter table, take out the grains at the surface layer, middle layer and layer depth, crush the grains to a particle size, press them into tablets, evenly spread them in a quartz sample cup to make a near-infrared spectroscopy sample, then grind the grains to a particle size, press them again, and cover the surface with a polyester film to make an X-ray fluorescence sample;

[0032] Through near-infrared spectroscopy combined with the PLS regression model, invert the moisture and protein content in the near-infrared spectroscopy sample in the quartz sample cup, use an X-ray fluorescence analyzer combined with the FP spectrum decomposition method to quantitatively detect the heavy metal content, and generate a quantitative analysis result of the moisture, protein and heavy metal content in the grain pile;

[0033] Use the constrained Kriging interpolation method to spatially align the analysis results of the moisture, protein and heavy metal content in the grain pile with the three-dimensional density distribution map, generate a quality safety three-dimensional heat map integrating multiple physical fields, and calculate the quality safety indicators according to the kernel function combined with the density gradient, temperature and humidity.

[0034] In the second aspect, the present invention provides a grain depot intelligent sampling quality inspection system, including a data collection module that collects environmental data, three-dimensional point cloud data and surface acoustic wave group velocity data, and uses the Love wave equation to perform inversion calculation to obtain a three-dimensional density distribution map;

[0035] A sampling parameter module that calculates the vertical density gradient according to the three-dimensional density distribution map to obtain a stratification sampling parameter table;

[0036] A sample distribution detection module that performs stratified puncture on the grain pile according to the stratification sampling parameter table to generate actual sampling depth data, and uses a three-way sample valve to perform sample distribution detection on the actual sampling depth data;

[0037] The quality and safety module uses near-infrared spectroscopy and X-ray fluorescence to analyze the moisture, protein, and heavy metal content in the grain pile, matches the analysis results with the density distribution map, and obtains the quality and safety indicators.

[0038] In a third aspect, the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and: when the computer program is executed by the processor, any step of the intelligent sampling and quality inspection system and method for a grain depot as described in the first aspect of the present invention is implemented.

[0039] In a fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and: when the computer program is executed by the processor, any step of the intelligent sampling and quality inspection system and method for a grain depot as described in the first aspect of the present invention is implemented.

[0040] The beneficial effects of the present invention are as follows: Through the inversion of the hexagonal SAW sensor array and the Love wave equation, the present invention realizes a high-precision mapping of the three-dimensional density distribution of the grain pile, with a spatial resolution of 15 cm and an inversion error ≤ 3.5%. It solves the deficiencies of traditional detection methods in blind area coverage and density decoupling. Combining the dynamic piercing mechanism of the Terfenol-D magnetostrictive actuator, the sampling path is optimized by the maximum heap algorithm, and the hard caking layer is penetrated intelligently with adjustable speed to improve the piercing efficiency. Through the hierarchical detection of the three-way sampling valve and the combined analysis of near-infrared and X-ray, the accurate detection of moisture, protein, and heavy metal content is realized. By using the constrained Kriging interpolation method to match the detection results with the density distribution map, quality and safety indicators are generated. Description of the Drawings

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0042] Figure 1 It is the intelligent sampling and quality inspection flow chart for the grain depot in Embodiment 1;

[0043] Figure 2 It is the data flow chart of the intelligent sampling and quality inspection system for the grain depot in Embodiment 1;

[0044] Figure 3 It is the intelligent sampling and quality inspection system diagram for the grain depot in Embodiment 1;

[0045] Figure 4 It is the module and process schematic diagram of the intelligent sampling and quality inspection system for the grain depot in Embodiment 1. Detailed Embodiments

[0046] To make the above objects, features, and advantages of the present invention more apparent and understandable, the following provides a detailed description of the specific embodiments of the present invention in conjunction with the accompanying drawings of the specification.

[0047] In the following description, many specific details are set forth in order to provide a thorough understanding of the present invention. However, the present invention may be implemented in other ways different from those described herein. Those skilled in the art can make similar generalizations without departing from the spirit of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0048] Secondly, the so-called "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation manner of the present invention. The appearances of "in one embodiment" in different places in this specification do not all refer to the same embodiment, nor are they separate or alternative embodiments that exclude each other from other embodiments.

[0049] Example 1, referring to Figure 1 、 Figure 2 、 Figure 3 and Figure 4 , is the first embodiment of the present invention. This embodiment provides an intelligent sampling and quality inspection system and method for a grain depot, including the following steps:

[0050] S1. Collect environmental data, three-dimensional point cloud data, and surface acoustic wave group velocity data.

[0051] S1.1. Scan the surface of the grain pile with a lidar to generate three-dimensional point cloud data.

[0052] Furthermore, based on the RIEGL VZ-4000 type three-dimensional lidar, a dynamic scan is constructed, with a 5 mm point spacing and a 10 Hz repetition frequency scan mode. The scattered noise of the particles on the surface of the grain pile is eliminated through an adaptive surface fitting algorithm to generate structured point cloud data containing a normal vector field. The scanning process achieves millimeter-level positioning accuracy through GNSS / INS combined navigation, and three-dimensional point cloud data is generated in cooperation with a point cloud registration algorithm.

[0053] S1.2. Arrange a 64-channel SAW sensor array in a hexagonal distribution on the surface of the grain pile. By emitting surface acoustic waves in the range of 50 - 150 kHz and measuring the propagation time difference between adjacent sensors, the surface acoustic wave group velocity data is obtained.

[0054] Furthermore, a 64-element piezoelectric ceramic SAW sensing array is arranged in a hexagonal close-packed topological structure, and the node spacing is set to 30 cm according to the wavelength compression criterion. A linear frequency modulation signal of 50 - 150 kHz is generated by an AD9833 signal generator, and waveform data is captured by an AD7768 high-speed acquisition module at a sampling rate of 1 MS / s. The propagation time difference between adjacent nodes (with an accuracy of 0.1 μs) is calculated based on the cross-correlation algorithm to obtain the surface acoustic wave group velocity data.

[0055] S1.3. Collect environmental data including humidity and temperature.

[0056] Furthermore, the humidity parameters of the surface layer and deep layer (depth interval of 1 m) of the grain pile are collected by an SHT85 temperature and humidity sensor at a frequency of 1 Hz, and the temperature data is collected layer by layer at a sampling rate of 1 Hz by a distributed SHT85 digital temperature and humidity sensor (with an accuracy of ±0.1 °C).

[0057] S2. Use the Love wave equation for inversion calculation to obtain a three-dimensional density distribution map.

[0058] S2.1. Based on the obtained surface acoustic wave group velocity data, use the Love wave equation to invert the density of the grain pile and generate a three-dimensional density distribution, with the expression

[0059]

[0060] where is the second derivative of the displacement u in the vertical direction, is the second derivative at time t, is the infinitesimal change of the second derivative in the vertical direction, is the group velocity, C is the humidity, T is the temperature, and g is the identifier of the group velocity.

[0061] Furthermore, is usually a real number and can be positive, negative, or zero, is usually a real number, is not equal to 0 and represents a small change in spatial position, is usually a real number is not equal to zero and represents a small change in time, is usually positive, the range of C is 0 ≤ C ≤ 100%, T is positive, with the unit of Kelvin (K) or Celsius (°C), and g is a constant;

[0062] Adopt a full-waveform inversion framework, construct a density sensitivity kernel function based on the adjoint state method, and perform iterative solution on a voxel scale of 0.05 m through a GPU-accelerated conjugate gradient algorithm to finally generate a three-dimensional distribution map containing density anomaly regions. 3 Finally, a three-dimensional distribution map containing density anomaly regions is generated.

[0063] S2.2. Combine the three-dimensional density distribution with the three-dimensional point cloud data using the grid interpolation combination method to obtain a three-dimensional density distribution map.

[0064] Furthermore, based on the multi-scale data fusion framework, an improved unstructured grid interpolation algorithm is used to achieve sub-voxel level registration of the density field and the point cloud. First, the density inversion grid (5 cm resolution) and the lidar point cloud (2 mm accuracy) are spatially aligned through the iterative closest point (ICP) algorithm to obtain a three-dimensional density distribution map.

[0065] S3. Calculate the density gradient in the vertical direction according to the three-dimensional density distribution map to obtain a stratified sampling parameter table.

[0066] S3.1. Align the density data with the lidar point cloud coordinates through the grid alignment method, and use the central difference method to calculate the density gradient layer by layer in the vertical direction of the grain pile.

[0067] Furthermore, the expression for calculating the density gradient layer by layer in the vertical direction of the grain pile is:

[0068]

[0069] where is the first derivative of the density gradient ρ in the vertical direction, z i is the space at the depth i in the vertical direction in the grain pile, Δz is the space step in the vertical direction, and i is the grain pile depth index;

[0070] is usually positive and greater than zero, z i is usually positive and greater than zero, Δz is usually positive and greater than zero, and i is usually an integer.

[0071] S3.2. Set the humidity mutation layer threshold based on the density gradient layer by layer in the vertical direction of the grain pile. When the density gradient is greater than the humidity mutation layer threshold, divide the boundary edge of the grain pile.

[0072] Furthermore, along the height direction of the grain pile, measure the density layer by layer, and set a density mutation layer threshold, which is the key criterion for judging whether there is a significant density change between adjacent layers. When it is monitored that the density gradient exceeds the preset density mutation layer threshold, it means that there is an obvious humidity difference between these two layers, and then these two layers can be divided into different grain pile boundary edges.

[0073] S3.3. Divide the grain pile into N layers according to the grain pile boundary edge to obtain a stratified sampling grain pile parameter table.

[0074] Further, according to the boundary edge of the grain pile determined by density gradient analysis, the entire grain pile is accurately divided into N different layers in the vertical direction. Each layer represents a region with relatively consistent density and humidity conditions. Detailed sampling work is carried out for these N layers respectively to form a stratified sampling grain pile parameter table. This parameter table details the specific measurement data of each layer.

[0075] S4. Use the stratified sampling parameter table to perform stratified punctures on the grain pile to generate actual sampling depth data.

[0076] S4.1. Perform priority sorting on the stratified sampling parameter table through the max heap data structure, dynamically analyze high-risk areas, and dynamically adjust the puncture speed in descending order of density gradient intensity.

[0077] Further, adopt the max heap dynamic sorting method to analyze the density gradient intensity in the stratified sampling grain pile parameter table, and dynamically adjust the puncture speed of the Terfenol-D magnetostrictive actuator;

[0078] Arrange in descending order according to the magnitude of the density gradient of each layer. For the layers with a larger density gradient, since there may be more complex structures or state changes, the puncture speed will be automatically slowed down to improve the accuracy of data collection; while for the layers with a smaller density gradient, the puncture speed will be appropriately increased to improve efficiency.

[0079] S4.2. Use the Terfenol-D magnetostrictive actuator to vertically penetrate the grain pile at the dynamically adjusted puncture speed, and record the actual sampling depth data in real time.

[0080] Further, based on the dynamically adjusted puncture speed, the Terfenol-D magnetostrictive actuator will vertically penetrate the corresponding layer of the grain pile at the optimal speed. During this process, the specific position and puncture depth of the actuator are monitored and recorded in real time to generate actual sampling depth data;

[0081] Through the Terfenol-D magnetostrictive actuator, it can ensure that each puncture accurately reaches the predetermined depth, and even in the face of a grain pile with uneven density, it can maintain a high degree of accuracy. These actual sampling depth data verify the effectiveness of the preset stratification.

[0082] S5. Use a three-way sampling valve to perform sample distribution detection on the actual sampling depth data.

[0083] S5.1. Set the surface layer humidity threshold and deep layer humidity threshold of the grain pile according to the overall humidity of the grain pile collected.

[0084] Further, according to the overall humidity of the grain pile collected, use the sliding window method to obtain the humidity distribution of the surface layer (0 - 0.5m) and the deep layer (>3m).

[0085] S5.2 Use a three-way sampling valve to conduct sample distribution detection on the actual sampling depth data.

[0086] Furthermore, the angle of the valve core is real-time feedback through a three-way sample splitting valve and an integrated Hall encoder (accuracy ±0.05°). According to the depth data of the pressure sensor (range 0 - 50 kPa, accuracy ±0.1 kPa) built into the sampling rod, the stepping motor (step angle 1.8°) is driven through the PID control algorithm to switch the flow channel. Shallow samples (<0.5 m) enter the mold detection channel (flow rate 2 L / min), middle-layer samples (0.5 - 3 m) flow to the moisture detection chamber (flow rate 1.5 L / min), and deep samples (>3 m) are introduced into the heavy metal detection module (flow rate 0.8 L / min). A self-cleaning flow channel design (compressed air pulse period 30 s) is adopted to avoid cross-contamination, and the sample splitting error rate <0.8%.

[0087] S5.3 Conduct mold detection when the actual sampling depth data is less than the surface humidity threshold.

[0088] Furthermore, the surface samples are subjected to high-throughput mold detection through a microfluidic chip (channel width 200 μm). The quantitative detection of the colony count (sensitivity 10 2 CFU / g) is achieved by integrating qPCR (temperature control accuracy ±0.3 °C) to amplify the aflatoxin synthesis gene and combining with a fluorescent probe (excitation wavelength 485 nm) to obtain the mold detection result. When the toxin concentration > 5 μg / kg, a red warning signal is triggered.

[0089] S5.4 When the surface humidity threshold is less than or equal to the actual sampling depth data and the actual sampling depth data is less than or equal to the deep humidity threshold, conduct moisture detection.

[0090] Furthermore, the middle-layer samples are rapidly detected for moisture content by using near-infrared spectroscopy (wavelength 900 - 1700 nm, resolution 3 nm) combined with a PLS regression model (number of principal components = 8). The spectral probe is equipped with a self-focusing lens (focal length 15 mm) and enters the grating spectrometer (integration time 100 ms) through optical fiber coupling. The detection accuracy reaches ±0.15%. The data is filtered by Kalman filter to eliminate the vibration noise during transportation, and the three-dimensional distribution heat map of the moisture content is updated every 60 seconds. Samples with excessive moisture content (moisture content > 14.5%) are automatically marked with yellow warning labels.

[0091] S5.5 When the actual sampling depth data is greater than the deep humidity threshold, conduct heavy metal detection.

[0092] Furthermore, after deep samples are digested by microwave digestion (power 1200W, heating rate 15°C / s), the contents of heavy metals such as lead and cadmium are detected by ICP-MS (argon plasma temperature 8000K). A collision reaction cell (He flow rate 4.3 mL / min) is equipped to eliminate mass spectrometry interference. The detection limit is as low as 0.01 μg / kg (RSD<5%), and the data is uploaded to the quality traceability system through the Modbus protocol. When Pb>0.2 mg / kg or Cd>0.1 mg / kg, a hierarchical control mechanism (isolation, ventilation, alarm) is triggered

[0093] S6. Near-infrared spectroscopy and X-ray fluorescence are used to analyze the moisture, protein, and heavy metal content in the grain pile. The analysis results are matched with the density distribution map to obtain quality safety indicators

[0094] S6.1. Based on the hierarchical parameter table, grains from the surface layer, middle layer, and deep layer are taken out, crushed to a particle size, tableted, and evenly spread on a quartz sample cup to make a near-infrared spectroscopy sample. Then the grains are ground to a particle size and tableted again, and a polyester film is covered on the surface to make an X-ray fluorescence sample

[0095] Furthermore, based on the hierarchical parameter table, representative grain samples are taken out from the surface layer, middle layer, and deep layer respectively, crushed to a suitable particle size, then tableted and evenly laid on a quartz sample cup to prepare a sample suitable for near-infrared spectroscopy analysis. The same or different grain samples are further ground to a finer particle size, tableted again and covered with a polyester film on the surface to prepare a sample suitable for X-ray fluorescence analysis

[0096] S6.2. By combining near-infrared spectroscopy with the PLS regression model, the moisture and protein content in the near-infrared spectroscopy sample in the quartz sample cup are inverted. Using an X-ray fluorescence analyzer combined with the FP spectral decomposition method, the heavy metal content is quantitatively detected to generate a quantitative analysis result of the moisture, protein, and heavy metal content in the grain pile

[0097] Furthermore, a near-infrared spectrometer is used to detect the samples placed in the quartz sample cup to determine the moisture and protein content therein, while an X-ray fluorescence analyzer can detect the heavy metal content in the samples, so as to obtain detailed information about the moisture, protein, and potentially harmful heavy metal content in the grain pile

[0098] S6.3. The constraint Kriging interpolation method is used to spatially align the analysis results of the moisture, protein, and heavy metal content in the grain pile with the three-dimensional density distribution map to generate a three-dimensional quality safety thermal map integrating multiple physical fields. The quality safety indicators are calculated according to the kernel function combined with the density gradient, temperature, and humidity

[0099] Furthermore, calculate the quality safety indicator, and the expression is

[0100]

[0101] Among them, Q(x) is the quality safety index at position x in the grain pile, and λ i is the moisture content weight coefficient at the i-th depth of the grain pile, W i is the moisture content at the i-th depth of the grain pile, α is the weight coefficient of the protein content, and μ i is the weight coefficient of the protein content at the i-th depth of the grain pile, P i is the protein content value at the i-th depth of the grain pile, β is the weight coefficient of the heavy metal content, and γ i is the weight coefficient of the heavy metal content at the i-th depth of the grain pile, H i is the heavy metal content value at the i-th depth of the grain pile, o(x) is the density value at position x, and o crit is the critical density, K is the kernel function, h is the sample, n is the total number of samples, and x is the position index in the grain pile;

[0102] λ i is usually a real number, and λ i ≥0, W i is usually a real number, and it can be positive, negative, or zero. α is usually a real number, and α≥0. μ i is usually a real number, and μ i ≥0, P i is usually a real number, and it can be positive, negative, or zero. β is usually a real number, and β≥0. γ i is usually a real number, and γ i ≥0, H i is usually a real number, and it can be positive, negative, or zero. o(x) is usually a real number, and it can be positive, negative, or zero. o crit is usually a positive value, and o crit is greater than zero. h is usually a real number and is greater than zero.

[0103] The constrained Kriging interpolation method is adopted to predict the data values of unknown points, which involves matching the analysis results of moisture, protein, and heavy metal contents with the density distribution map of the grain pile. Through effective processing of spatial data, a comprehensive quality safety index is finally generated.

[0104] Furthermore, the constrained Kriging interpolation method is adopted. By combining the analysis results of moisture, protein, and heavy metal contents with the three-dimensional density distribution map of the grain pile, the spatial data is refined to predict the data values of unknown points, thereby achieving a high-precision mapping of the internal quality characteristics of the grain pile. By introducing constraint conditions, the accuracy of the interpolation results is optimized to ensure a high degree of matching between the spatial distributions of moisture, protein, and heavy metal contents and the density distribution map. A comprehensive quality safety index is generated by weighted fusion of multi-dimensional data.

[0105] This embodiment also provides an intelligent sampling and quality inspection system for a grain depot, including: a data collection module that collects environmental data, three-dimensional point cloud data, and surface acoustic wave group velocity data, and performs inversion calculation using the Love wave equation to obtain a three-dimensional density distribution map;

[0106] A sampling parameter module that calculates the density gradient in the vertical direction based on the three-dimensional density distribution map to obtain a stratified sampling parameter table;

[0107] A sample allocation and detection module that performs stratified piercing on the grain pile according to the stratified sampling parameter table to generate actual sampling depth data, and uses a three-way sampling valve to perform sample allocation and detection on the actual sampling depth data;

[0108] A quality and safety module that uses near-infrared spectroscopy and X-ray fluorescence to analyze the moisture, protein, and heavy metal content in the grain pile, and matches the analysis results with the density distribution map to obtain quality and safety indicators.

[0109] This embodiment also provides a computer device applicable to the intelligent sampling and quality inspection system and method for a grain depot, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the intelligent sampling and quality inspection system and method for a grain depot as proposed in the above embodiment.

[0110] This computer device can be a terminal, and this computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Among them, the processor of this computer device is used to provide computing and control capabilities. The memory of this computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of this computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a carrier network, NFC (near field communication), or other technologies. The display screen of this computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of this computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the outer shell of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0111] This embodiment also provides a storage medium, on which a computer program is stored. When the program is executed by a processor, it implements the grain depot intelligent sampling quality inspection system and method proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, abbreviated as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, abbreviated as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, abbreviated as EPROM), programmable read-only memory (Programmable Red-Only Memory, abbreviated as PROM), read-only memory (Read-Only Memory, abbreviated as ROM), magnetic memory, flash memory, magnetic disk or optical disc.

[0112] In summary, through the hexagonal SAW sensor array and Love wave equation inversion, the present invention realizes a high-precision mapping of the three-dimensional density distribution of the grain pile, with a spatial resolution of 15 cm and an inversion error ≤ 3.5%. It solves the deficiencies of traditional detection methods in blind area coverage and density decoupling. Combining with the dynamic piercing mechanism of the Terfenol-D magnetostrictive actuator, it optimizes the sampling path through the maximum heap algorithm, intelligently adjusts the speed to penetrate the hard caking layer, and improves the piercing efficiency. Through the hierarchical detection of the three-way sampling valve and the combined analysis of near-infrared X-ray, it realizes the accurate detection of moisture, protein and heavy metal content. By matching the detection results with the density distribution map through the constrained Kriging interpolation method, a quality safety index is generated.

[0113] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A grain depot intelligent sampling quality inspection method, characterized by: include, Collect environmental data, three-dimensional point cloud data and surface acoustic wave group velocity data, use Love wave equation inversion calculation to obtain a three-dimensional density distribution map; Calculate the density gradient in the vertical direction according to the three-dimensional density distribution map to obtain a stratified sampling parameter table; Puncture the grain pile in layers according to the layered sampling parameter table to generate actual sampling depth data, and use a three-way sampling valve to perform sample distribution detection on the actual sampling depth data; Near-infrared spectroscopy and X-ray fluorescence were used to analyze the moisture, protein and heavy metal content in the grain pile. The analysis results were matched with the three-dimensional density distribution map to obtain quality and safety indicators.

2. The intelligent sampling quality inspection method for grain depot according to claim 1 is characterized in that: Collecting environmental data, 3D point cloud data and surface acoustic wave group velocity data includes the following steps: Scan the grain pile surface with LiDAR to generate 3D point cloud data; In the sensor array on the grain pile surface, the surface acoustic wave group velocity data is obtained by emitting surface acoustic waves and measuring the propagation time difference between adjacent sensors.

3. The intelligent sampling quality inspection method for grain depot according to claim 2 is characterized by: Using the Love wave equation to invert the calculation, the three-dimensional density distribution map includes the following steps: Based on the surface acoustic wave group velocity data, the grain bulk density is inverted using the Love wave equation to generate a three-dimensional density distribution. The three-dimensional density distribution is combined with the three-dimensional point cloud data using the grid interpolation method to obtain a three-dimensional density distribution map.

4. The intelligent sampling quality inspection method for grain depot according to claim 3 is characterized by: Calculating the vertical density gradient according to the three-dimensional density distribution map to obtain the stratified sampling parameter table includes the following steps: The density data is aligned with the laser radar point cloud coordinates by grid alignment method, and the central difference method is used to calculate the density gradient of each layer in the vertical direction of the grain pile. The density mutation layer threshold is set based on the density gradient layer by layer in the vertical direction of the grain pile. When the density gradient is greater than the density mutation layer threshold, the grain pile boundary is divided. The grain pile is divided into N layers according to the boundary edges of the grain pile, and a stratified sampling grain pile parameter table is obtained.

5. The intelligent sampling quality inspection method for grain depot according to claim 4 is characterized in that: Using the stratified sampling parameter table to puncture the grain pile in layers and generate actual sampling depth data includes the following steps: The stratified sampling parameter table is prioritized through the maximum heap data structure, high-risk areas are dynamically parsed, and the puncture speed is dynamically adjusted according to the descending order of density gradient intensity; The Terfenol-D magnetostrictive actuator was used to vertically penetrate the grain pile at a dynamically adjusted penetration speed, and the actual sampling depth data was recorded in real time.

6. The intelligent sampling quality inspection method for grain depots according to claim 5 is characterized by: The sample distribution test of the actual sampling depth data using a three-way sampling valve includes the following steps: According to the overall humidity of the grain pile, the surface humidity threshold and deep humidity threshold of the grain pile are set; Use a three-way sampling valve to perform sample distribution detection on the actual sampling depth data; When the actual sampling depth data is less than the surface humidity threshold, mold detection is carried out; When the surface humidity threshold is less than or equal to the actual sampling depth data, and the actual sampling depth data is less than or equal to the deep humidity threshold, moisture detection is performed; When the actual sampling depth data is greater than the deep humidity threshold, heavy metal detection is performed.

7. The intelligent sampling quality inspection method for grain depot according to claim 6 is characterized by: The moisture, protein and heavy metal content in the grain pile is analyzed by near infrared spectroscopy and X-ray fluorescence, and the analysis results are matched with the density distribution map to obtain the quality and safety indicators. The following steps are included: Based on the stratification parameter table, the grains from the surface layer, middle layer and layer depth were taken out, the grains were crushed to particle size, pressed into tablets, and evenly spread in a quartz sample cup to make near-infrared spectroscopy samples, and then the grains were ground to particle size, pressed into tablets, and the surface was covered with polyester film to make X-ray fluorescence samples; The moisture and protein content in the near-infrared spectrum samples in the quartz sample cup were inverted by combining near-infrared spectroscopy with the PLS regression model. The heavy metal content was quantitatively detected using an X-ray fluorescence analyzer combined with the FP spectrum analysis method to generate quantitative analysis results of the moisture, protein and heavy metal content of the grain pile. The constrained Kriging interpolation method is used to spatially align the analysis results of moisture, protein and heavy metal content in the grain pile with the three-dimensional density distribution map to generate a three-dimensional heat map of quality and safety that integrates multiple physical fields. The quality and safety indicators are calculated based on the kernel function combined with density gradient, temperature and humidity.

8. An intelligent sampling and quality inspection system for grain depots, based on the intelligent sampling and quality inspection method for grain depots according to any one of claims 1 to 7, characterized in that: include, The data collection module collects environmental data, three-dimensional point cloud data and surface acoustic wave group velocity data, and uses the Love wave equation for inversion calculation to obtain a three-dimensional density distribution map; The sampling parameter module calculates the vertical density gradient according to the three-dimensional density distribution map to obtain a layered sampling parameter table; The sample distribution detection module punctures the grain pile in layers according to the layered sampling parameter table, generates actual sampling depth data, and uses a three-way sampling valve to perform sample distribution detection on the actual sampling depth data; The quality and safety module uses near-infrared spectroscopy and X-ray fluorescence to analyze the moisture, protein and heavy metal content in the grain pile, and matches the analysis results with the density distribution map to obtain quality and safety indicators.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the intelligent grain storage sampling and quality inspection method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the intelligent grain storage sampling and quality inspection method described in any one of claims 1 to 7 are implemented.

Images